基于梯度相位和显著性约束的Hausdorff 距离模板匹配方法

针对复杂背景下形状不规则、高度较低的平面目标自动识别问题,提出了一种基于Hausdorff距离的模板匹配方法。在完成平面目标前视模板制备后,文中首先定义了基于边缘位置、梯度相位和边缘点显著性约束的相似性度量方法,模板与实时图中对应两个边缘点位置越近、梯度相位差越小及实时图边缘点越显著,这两点的匹配就越好;然后融合三种度量结果,设计了一种基于边缘相位和显著性约束的Hausdorff 距离模板匹配方法,实现了平面目标轮廓的准确匹配。实测数据处理结果表明,该方法能够实现复杂地面场景中任意形状的平面目标轮廓的匹配定位,并且定位精度高、鲁棒性好、适用范围广

Hausdorff distance template matching method based on gradient phase and significance constraints

Aiming at the problem of automatic recognition of irregular shaped and lower height plane targets under the complex background, a template matching method based on Hausdorff distance was proposed. After the forward looking template of plane target was prepared, a similarity measurement method was defined based on edge position, gradient phase and edge point significance constraints firstly. When the positions of two edge points in template and real-time image were close, the gradient phase difference was small and the edge point in real-time image was significant, the two points were matched to each other; then a kind of Phase and Confidence Restrict Hausdorff Distance(PCR-HD) template matching method was designed through the fusion of above three metrics, which achieved the exact matching of plane target contour. The processing results of measured data show that, the proposed algorithm can realize contour matching and recognition of plane target with arbitrary shape under complex ground scene, and has high location accuracy, good robustness and wide application range.